Skip to content

Who builds it

The Marella AI team

The people who build Marella AI, from the interface design to the AI systems underneath, and how the team decides what the product will claim.

Updated 21 Aug 2026

01

Leadership

Ibrahim Mizi is co-founder and CEO. He holds a first-class Computer Science degree from Durham University, works mainly in frontend design with React and Vue.js, and builds language-model features into the interfaces he designs.

Reuben McQueen is co-founder and CTO. He holds a first-class Computer Science degree and works across system architecture, automation, real-time data systems and frontend development, applying large language models, vision language models and other multimodal systems to business problems.

  • Ibrahim Mizi leads product, interface and company direction
  • Reuben McQueen leads architecture, AI systems and engineering
  • Frontend design and AI-assisted interfaces
  • Automation, real-time systems and cloud development
02

Engineering

Flynn holds a PhD in Biomedical Engineering from the University of Strathclyde and has delivered AI systems in education, healthcare and finance. He designs and deploys LLM pipelines covering prompting, retrieval, fine-tuning and evaluation, along with the multi-agent and human-in-the-loop architectures around them.

Margarita works in React and TypeScript with both SQL and NoSQL databases, including Supabase and MongoDB, and designs interfaces people can find their way around without being taught.

  • Flynn Lachendro works on AI systems and evaluation
  • Margarita Marmaridou works on full-stack product engineering
  • LLM pipelines, multi-agent orchestration and human-in-the-loop design
  • React, TypeScript and SQL or NoSQL data work
03

How the team works

We reconcile every product claim against the implementation behind it, and we keep evaluation work close to the documents and the reviewers involved.

  • Claims reconciled with implementation
  • Evaluation close to the documents
  • Reviewers kept involved

Test the claim on your documents

Pick a real piece of work, agree what a good answer looks like, then go through the results together.